Vision-based Control System Design For Mini Underwater Vehicle

Underwater robot is an important carrier in the work of complex marine environment.With the growing importance of marine resources development, underwater robots have been used more and more in marine mineral exploration,seabed topography and deep sea probe inspection infrastructure. Underwater optical vision can provide richer information in the underwater robot close observation operations, attracted more and more researchers' attention from all over the world. Therefore, the research on the control system of the mini underwater vehicle based on vision has important theoretical significance and practical application value for improving the autonomous ability and intelligent level of the underwater vehicle.In the study of visual target detection method of underwater vehicle. Firstly, in order to solve the problem of the underwater image degradation, so as to establish the underwater light propagation attenuation model and underwater imaging system model, through the aqueous medium in the attenuation of light color recovery algorithms to recover complete image preprocessing; Then,an adaptive color image segmentation method based on HSV is used in the segmentation algorithm of the underwater object, the extraction of moving object region is extracted by the method of constraint extraction of regional geometric shape feature parameters. Finally, the study of moving target tracking method, using based on Kalman predictor combined with dynamic detection window of target tracking algorithm, the prediction results are corrected by extracting the centroid information of the image, and the experimental study of moving target tracking in underwater vehicle is carried out in order to verify the real time and effectiveness of the target tracking method.Research and design the underwater vehicle control system based on the vision target recognition and tracking system. Firstly, the overall structure of the control system and its communication bus is designed. Then, each subsystem of the distributed system is designed in detail, including the host computer system, communication system, thruster control system,power management and security system and sensor information acquisition system. And the software structure of the control system is designed, and finally the distributed control system based on vision is formed by the software and hardware system.In the study of motion control system simulation of underwater vehicle. Firstly, the design of mini underwater vehicle hull structure and the distribution of the sensors are introduced; Then, the six degree of freedom motion model of underwater vehicle is analyzed,and decoupling the dynamic equations of expression of heave and heading. Finally, the decoupling control system and the lifting subsystem are designed and simulated, which is the theoretical groundwork for the underwater target identification and tracking control experiment.Finally, a large number of experiments are conducted on the small underwater vehicle,including the sensor filtering experiment, depth and heading control experiment, combined with the target recognition based on visual tracking system design for autonomous underwater target tracking experiment, to verify the real-time performance and stability of the control system designed in this paper.